Page 194 - Artificial Intelligence for the Internet of Everything

180 Artificial Intelligence for the Internet of Everything

replace them. However, in each of these aggressive steps, there are not only
promises, but also pitfalls, which, if outweighed, can together paralyze the
design of autonomous agents as human incarnations. In what follows we
shall provide a short, but not exhaustive, list of impediments. The three
points that we shall raise below are all concerned with the economics of atten-
tion (Lanham, 2006).
First, IoE has endowed us with a degree of information richness to an
unprecedented level. Now, within the IoE environment, information is
constantly flowing around us. Some of our idle time, such as waiting for
the bus, the traffic light, the elevator, or the waiter, is now recruited to pro-
cessing the constant inflow of information. We then gradually lose these idle
moments, and certainly cannot entertain them with our conventional styles
of dazing, dreaming, pondering, reflecting, or run-away thinking. A neol-
ogism, the smartphone zombie, has been invented to describe this unparalleled
phenomenon. While the effect of IoE on our brains and minds is unclear at
this point, it may be fair to say that it has shaped or defined what thinking can
mean for us. Now the conventional way of thinking by “staring at the wall”
or offline thinking is being dramatically replaced by online thinking, i.e., think-
ing with constant interruptions. Comparing these two different styles of
thinking is not that straightforward, and while the familiar dual-task or mul-
tiple-task psychological experiments may help shed some light on this com-
parative study, further work needs to be done (Damos, 1991). However,
when decision makers are often required to make their decisions online
and spontaneously, the deliberation efforts normally required by homo eco-
nomicus may be severely curtailed.
Second, a related point is that our attention capacity is a scarce resource
and the information richness brought by IoE can make attention scarcity
even more strained. In the literature this tension is known as information over-
load (Sutcliffe & Weick, 2008). Information overload can adversely affect
our decision-making capability, which has already been well studied in psy-
chology and behavioral economics (Iyengar, 2010). In fact, Herbert Simon
(1916–2001) had long noticed such a tension (Simon, 1971). Although he
did not immediately exclude the possibility that the advances in ICT may
compound the information overload problem rather than solve or mitigate
it, he did provide the condition required for avoiding the appearance of the
downside.
“Whether a computer will contribute to the solution of an information-
overload problem or instead compound it depends on the distribution of its